Currently, in many application scenes, each function of the application needs to be used so as to ensure that files, network and other actions corresponding to the application can be triggered. For example, in an automatic virus detection process, when all of the actions can be triggered, it can be easily determined whether the application is a virus file; as another example, in an automatic garbage recognition process, when all of the files can be triggered, it can be recognized whether the application is a garbage file. However, these are ensured by requiring a complete click-trigger, and therefore a full automatic trigger is necessary.
For a present automatic click method, it generally divides a screen into different area zones according to pixel points of the screen, and then clicks the area zones one by one. However, this click method not only wastes time, but also causes much useless click, and it cannot ensure that all of the clickable widgets are clicked, such that the efficiency is lower and the accuracy is lower.